This invention relates to a technique for managing a plurality of I/O (input/output) paths between a storage and a server.
In recent years, a SAN (Storage Area Network) constructing a dedicated network between a storage (also called a storage subsystem) such as disk devices and a server has been closely watched. This construction is expected to reduce the storage management cost and make it possible to effectively utilize the storage capacity.
In the SAN environment, I/O requests are issued from a plurality of application programs at a time to a component element (array group, etc.) for I/O processing in the storage. In the case where the I/O processes of any one of the application programs increase, therefore, the I/O performance of the other application programs is deteriorated. This problem is generally called the I/O collision.
While the existing application program is operating on the server, assume that a new application program is assigned to the component element for I/O processing in the storage. It is necessary to take into account both the effect of I/O performance of the new application program due to the I/O process of the existing application program and the effect of I/O performance of the existing application program due to the I/O process of the new application program.
In this situation, a conventional method has been disclosed to avoid the I/O collision for an array group in the storage. According to this method, before selecting an array group for which a new volume is created, the performance requested for the new volume is acquired. Next, all the volumes associated with each array group to which the new volume can be assigned are checked whether they meet the requested performance. In the presence of any array group meeting the requested performance, the new volume is created for the particular array group (See U.S. Pat. No. 6,748,489 for example).